The Passivity And Pitting Of Copper Based On The PDM

Copper is the assumed material for the canisters in KBS-3repository concept for dis-posal of high-level nuclear waste in Sweden and is intended to completely isolate the spent nuclear fuel from the surroundings until failure occurs of the5cm thick copper canister. The passivity of pure copper over a wide pH range has been explored using potentiody-namic and potentiostatic polarization, Mott-Schottky analysis, and electrochemical imped-ance spectroscopy (EIS). The study shows that the passive film is a p-type semiconductor and the natural logarithm of the passive current density is found to linearly increase with the film formation potential. According to the predictions of the Point Defect Model (PDM), the dominant point defect in the copper oxide passive film is the cation vacancy. The passive state of copper is interpreted by optimizing the PDM on the EIS data using a genetic algorithm approach. Kinetic Stability Diagrams (KSDs), which are proposed as kinetic alternatives to the equilibrium thermodynamic Pourbaix diagrams, are derived for pure copper in borate buffer based on the rate law for passive film growth in accordance with the PDM.Passivity breakdown and pitting on pure copper, has been studied in simulated, sul-phide-and chloride-containing granitic rock groundwater and the data are interpreted in terms of the PDM. The results indicated that a sulphide barrier layer forms on the copper surface, and that chloride ion can induce passivity breakdown, and hence induce the devel-opment of localized corrosion damage (LCD), such as pitting corrosion, on the canister surface. In this study, the effects of chloride concentration (0.01-5mol/L), solution pH (=8～10), and temperature (22～82℃) on the pitting corrosion of copper in simulated repos-itory ground water have been studied. The principal findings of the study are as follows:Pitting corrosion is observed on copper, particularly at high chloride concentration, low pH, and high temperature, in simulated sulphide-and chloride-containing granitic rock groundwater where cuprous sulphide (Cu2S) comprises the barrier layer of the passive film. The results show that increasing chloride concentration and temperature shift the mean breakdown potential in the negative direction, thereby rendering the metal more susceptible to pitting corrosion. Analysis of the dependence of the breakdown potential on chloride activity, pH, and potential scan rate using the relationships predicted by the Point Defect Model (PDM) yields the value for the polarizability (α=0.475～0.54) of the barrier lay-er/solution (f/s) interface at different temperatures, the value for the dependence of the po-tential drop across the f/s interface on the pH, β=-0.0109V, and the critical, areal cation vacancy concentration, ξ leading to passivity breakdown on copper, which is found to be in the range of1.04×1015～1.14×1015cm-2at different temperatures. This is in excellent agreement with that (1014～1015cm-2) estimated structurally, assuming vacancy condensa-tion on either the copper substrate or on the cation sublattice of the barrier layer at the m/f interface. Experimental results have demonstrated the validity of the PDM for describing passivity breakdown on copper in chloride-containing sodium sulphide solutions.The experimentally-determined, near normal distributions in the cumulative probabil-ity of the breakdown potential for different chloride concentrations and temperatures are found to be in satisfactory agreement with theory in the quantitative characterization of the breakdown potential distributions for Cu as calculated from the PDM. Because the statis-tical cumulative probability distributions approach the abscissa asymptotically as the po-tential is reduced, there exists a finite probability that passivity breakdown might occur at a potential that is considerably more negative (perhaps by200or300mV) than the mean breakdown voltage, particularly in the light of the large surface area of a canister and the excessively long exposure time in the repository (>105years). Such breakdown sites would be characterized by abnormally high cation vacancy diffusivities and are likely to be rare. However, metastable pitting events, which appear as "spikes" on the polarization curve, are observed at potentials that are more than500mV more negative than the mean critical breakdown voltage, indicating that such sites do exist. Should one or more of these pit nuclei stabilize via the establishment of differential aeration, then stable pitting will result and cause macroscopic damage to the canister. The pitting potential has been estimated for various times along the corrosion evolutionary path out to a horizon of10,000years by combining the predicted change in chloride concentration and temperature with the dependence of the pitting potential on the chloride concentration and temperature, as meas-ured in this work. The pitting potential is predicted to shift slightly in the negative direc-tion (pitting becomes easier) during the initial heat-up of the repository, followed by a long trend to more positive values (pitting becomes more difficult) as the temperature declines.Preliminary measurements have been made of the maximum pit depth from which the pitting factor is estimated, knowing the passive current density, and hence the general cor-rosion loss, just prior to pit nucleation. The passive current density is used to estimate the general corrosion metal loss using Faraday’s law. The pitting factor (PF) is found to de-crease exponentially with time, but it eventually levels off to a value between4.4and14.0, depending upon the potential, temperature and chloride concentration. The maximum pit depth is found to approach a constant value with increasing time, as has been observed for many other systems. The preliminary measurements of the repassivation potential show that this parameter decreases significantly with increasing charge passed during pit growth. These predictions are "preliminary", because we do not yet have values of high fidelity for all of the parameters that determine the rate of evolution of the damage. However, these preliminary studies of the accumulation of pitting damage do illuminate the path that must be taken in future studies. Damage Function Analysis (DFA), particularly by applying Deterministic Extreme Value Statistics (DEVS), can be used to predict the accumulation of damage as the canister transitions along the corrosion evolutionary path of the repository and, hence, a quantitative assessment may be made of the fate of the canister over105years into the future.